Neurotypica

Every context is packed with cues---the physical space, the people around you, the objects you see, the smells, the sounds---and those cues directly activate the behaviours you've practised there before. Walk into your office, and "work mode" loads automatically. Walk into your childhood home, and you might find yourself acting like you're sixteen again. The cues aren't neutral; they're triggers for input-output mappings you've trained hundreds or thousands of times in that exact context.

So what can you do? If you want new behaviour to stick, train it in the context where you'll actually need it. Practising calm breathing at home won't help much if you need it at work---the work context doesn't have the cues that retrieve that skill. Alternatively, design your context deliberately: rearrange furniture, change routes, add new objects that cue the intended behaviour. You're not fighting willpower; you're working with the system by making the context cue what you want.

Your brain builds context-specific predictive models: "In this place, with these people, I do this thing." When you return to a familiar context, the brain loads the model it built there and starts predicting---and then acting to fulfil those predictions. This is why rehab gains vanish at home: home cues the old model ("this is where I drink", "this is where I scroll endlessly"), and the brain acts to minimise prediction error by doing what it predicts.

So what can you do? You need to retrain the model in the actual context. If possible, make the home context unfamiliar enough that the old model doesn't load cleanly (rearrange the furniture, change your routine, alter the cues). Or practise the new behaviour in the home context repeatedly until the prediction updates. Moving to a new place works so well for change because the brain has no confident predictions there yet---everything is uncertain, so exploration is cheap and new patterns can establish without fighting the old model.

Familiar contexts make trained routines run automatically, and control only wakes up if there's conflict or accountability. In a new context (rehab, retreat, vacation), there's uncertainty and novelty, so control stays more engaged. Back home, the familiar context lets routines run on autopilot, and by the time control notices, the old behaviour has already executed.

So what can you do? Don't rely on willpower to fight familiar contexts. Instead, redesign the context so the routine you want is the default. Add prompts that preload your intention before the cue arrives. Build in accountability (tell someone, work in a shared space). Or create artificial novelty---even small changes to the environment can raise enough uncertainty that control stays engaged longer.